How to Clean Up Interview Recordings — A Step-by-Step Guide for Journalists and Podcasters

You sit down with a great guest, have a brilliant conversation, and feel good about what you recorded. Then you pull up the file and it's a mess — one voice is too quiet, the other is boomy, there's a dog barking in the background at the 12-minute mark, and the phone connection sounds like it was routed through 1994. Welcome to interview audio.

This is one of the most common challenges for podcasters and journalists alike. Unlike a solo recording where you control every variable, an interview means dealing with whatever your guest's environment and setup happens to be. The good news is that most interview audio problems are fixable — if you know what to do and in what order to do it.

Why Interview Audio Is Uniquely Challenging

There are a few things that make interview recordings harder to clean up than solo content. You're dealing with two voices that were recorded in completely different acoustic environments. The levels are almost never matched — one speaker is always louder. And if the interview was done over a phone or video call, one side often has codec compression artifacts that no amount of EQ can fully undo.

The most problematic interview scenarios are: remote calls recorded through Zoom or similar (one side compressed), phone interviews (narrowband audio, low-frequency cut already applied by the network), and outdoor or noisy location interviews where ambient noise is unavoidable.

Before the Interview: The Setup That Makes Cleanup Easier

The single most impactful thing you can do for interview audio quality happens before you press record. For remote interviews, ask your guest to use headphones — this prevents their speakers from bleeding back into their microphone and causing echo. Ask them to find a quiet room and close the door. If possible, have them record their side locally and send it to you afterward — a "double-ender" setup like this gives you two clean files to work with instead of one compressed call recording.

For in-person interviews, get the microphone as close to both speakers as is practical. A lav mic clipped to each person is the gold standard. A single recorder placed on a table in between rarely produces great results.

When You Get Bad Audio Back: What to Do

Sometimes you have no choice. A guest records themselves on their phone in a kitchen while the dishwasher is running. Or the call dropped twice and auto-reconnected with a different codec. Here's how to approach damage control:

First, listen through the whole recording and make notes on the specific problems: Is there a constant background noise? Are there specific loud events (doors, notifications, coughs)? Is one voice significantly quieter? Is there reverb? Knowing exactly what you're dealing with before you start processing saves time and avoids over-processing sections that don't need it.

Second, split the two voices onto separate tracks if you can. Even if you only have a mixed file, try to find moments where only one person is speaking and note the level difference. That tells you how much gain adjustment you need to balance them.

The Cleanup Workflow Step by Step

Step 1 — Noise reduction. Apply noise reduction to any track with a consistent background sound. Use your recorded room tone as the noise profile if you have one. Be conservative — aggressive noise reduction creates an unpleasant metallic, watery artefact sound that's often worse than the original noise. Aim to reduce the noise floor, not eliminate it entirely.

Step 2 — Normalise levels between speakers. Match the perceived loudness of both voices. Use your ears rather than just the meters — average loudness (LUFS) is a better guide than peak levels here. A good target is to have both speakers feel equally present in the mix.

Step 3 — Voice EQ if needed. For voices that sound boxy or muddy, a high-pass filter at around 80–100 Hz removes low-end rumble without affecting the voice itself. If a voice sounds thin and distant (common with phone recordings), a gentle boost around 1–3 kHz helps bring it forward.

Step 4 — Compression. Apply gentle compression to each voice track independently. This tightens up any remaining level inconsistencies and makes both voices feel controlled and present. Use a moderate ratio (around 3:1) and avoid squashing the dynamics completely — you want the conversation to still feel natural.

Step 5 — Final mix check. Play the whole thing back on earbuds or laptop speakers, not just studio headphones. This is how most of your audience will listen. If something jumps out as loud or inconsistent, fix it now rather than after publishing.

Using VoxBoost AI in Your Interview Workflow

If you're doing a live remote interview, open VoxBoost AI and run your microphone through it before the call starts. You'll get real-time noise gating, EQ, and compression on your side of the conversation, which means at least one of the two audio tracks will be clean regardless of what happens on the guest's end.

For recorded interviews, the noise reduction and EQ presets in VoxBoost AI are a fast starting point — particularly useful if you're on a deadline and don't have time to tweak settings from scratch. Upgrade to premium for access to the full processing chain and unlimited recording length.